AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 148:589–600 (2012)

Stable Isotope Evidence for Sex- and Status-Based Variations in Diet and Life History at Medieval Vercellese, Laurie J. Reitsema1,2* and Giuseppe Vercellotti2

1Department of Anthropology, University of Georgia, Athens, GA 30602 2Department of Anthropology, The Ohio State University, Columbus, OH 43210

KEY WORDS carbon; nitrogen; collagen; dentine; millet

ABSTRACT The medieval period in Europe was a life course among these groups by comparing dentine time of unprecedented social complexity that affected and bone collagen. Our results show a diet based on ter- human diet. The diets of certain subgroups—for exam- restrial resources with input from C4 plants, which ple, children, women, and the poor—are chronically could include proso and/or foxtail millet. Diets of low- underrepresented in historical sources from the medie- status males differ from those of females (both status val period. To better understand diet and the distribu- groups) and of high-status males. These differences de- tion of foods during the medieval period, we investi- velop in adulthood. Childhood diets are similar among gated stable carbon and nitrogen isotope ratios of 30 the subgroups, but sex- and status-based differences individuals from Trino Vercellese, (8th– appear in adulthood. We discuss the possibility of cul- 13th c.). Specifically, we examined diet differences tural buffering and dietary selectivity of females and between subgroups (males and females, and high- and high-status individuals. Am J Phys Anthropol 148:589– low-status individuals), and diet change throughout the 600, 2012. VC 2012 Wiley Periodicals, Inc.

During the medieval period in Europe, human popula- In recent years, valuable insights into dietary varia- tions were coping with biological and social stresses tions of the general medieval populace have been pro- including urbanism, climate change, growing socioeco- vided by stable isotope analyses of human skeletal nomic differentiation, and increased interconnectedness remains with known sex, age, and, in some cases, status. that facilitated disease transmission, interpersonal inter- Stable carbon and nitrogen isotope signatures measured actions, and access to nonlocal technologies and goods in bone collagen reflect a composite of the isotope signa- (White, 1962; Shahar, 1990; Storey, 1992; Dyer, 1994; tures of foods consumed during an individual’s lifetime Aberth, 2005). The medieval cultural environment and are a widely used tool in anthropology for recon- impacted human diet, facilitating access to nonlocal structing past human diet [for a recent review, see foods and food technologies and promoting food dispar- Schoeninger (2011)]. Stable isotope studies of medieval ities along social, economic, and religious hierarchies Europe have focused on diet change through time (e.g., laity and clergy, peasants and the elite, rural and (Mu¨ ldner and Richards, 2007), weaning (Richards et al., urban dwellers, and men and women; Nada Patrone, 2002; Fuller et al., 2003), sex-based differences in diet 1981; Montanari, 1988; Adamson, 2004). (Richards et al., 2006; Kjellstro¨m et al., 2009; Nitsch and Much of what is known about medieval diet comes Hedges, 2010; Reitsema et al., 2010), and status-based from historical accounts pertaining to the sociopolitical differences in diet (Schutkowski et al., 1999; Czermak and religious elite. Information about the general popu- et al., 2006; Kjellstro¨m et al., 2009; Mu¨ ldner et al., lace is less readily available, coming primarily from 2009). These studies underscore the dynamic ways in archaeology and what can be inferred from a limited which culture and circumstance affected food access number of historical accounts. These sources indicate that medieval food access followed strict delineations based on age, sex, and status (Adamson, 2004). However, Additional Supporting Information may be found in the online additional research is needed to shed light on the com- version of this article. plex local and transcontinental biocultural factors gov- erning food access and consumption in medieval society. Grant sponsor: Sigma Xi, the Scientific Research Foundation. Relatively large skeletal samples from the medieval pe- riod enable stable isotope studies to address research *Correspondence to: Laurie J. Reitsema, Department of Anthro- questions about diets of both populations and of individ- pology, The Ohio State University, 4034 Smith Laboratory, 174 W. 18th Avenue, Columbus, OH 43210, USA. uals, through time and between regions, including E-mail: [email protected]

 Did diets of high- and low-status individuals differ? Received 22 February 2012; accepted 29 March 2012  Were there sex-based differences in diet?  Did diet change during an individual’s lifetime? DOI 10.1002/ajpa.22085  How did diet impact health and fitness for various Published online 3 May 2012 in Wiley Online Library subpopulations? (wileyonlinelibrary.com).

VC 2012 WILEY PERIODICALS, INC. 590 L.J. REITSEMA AND G. VERCELLOTTI during the medieval period. Regional variation in diet Stable isotopes in anthropological diet during the medieval period is impressive, and more reconstructions research is needed to understand specific biocultural adaptations and overall dietary variation at both trans- Stable isotope analysis has become a widespread tech- continental and local scales. nique in anthropology for reconstructing past human Beyond basic diet reconstruction, an important contri- diet since its initial applications in the 1970s (Vogel and bution of stable isotope analysis is the reconstruction of van der Merwe, 1977; van der Merwe and Vogel, 1978). diet change over the life course using stable isotope sig- Stable carbon and nitrogen isotopes provide differing natures from different components of the skeleton. Previ- and complementary information about human diet ous stable isotope research has capitalized on the fact (Schoeninger, 2011). In general, stable carbon isotopes that different parts of the skeleton either form at differ- reflect the local ecosystem of a consumer and consump- ent ages (e.g., teeth vs. bones) or model and remodel at tion of aquatic versus terrestrial foods, whereas stable different rates (e.g., cortical vs. trabecular bone) and nitrogen isotopes provide information about an animal’s thus represent different time periods in an individual’s trophic position (Vogel and van der Merwe, 1977; Mina- life. With the entire human skeleton functioning as a waga and Wada, 1984; Schoeninger and DeNiro, 1984). storehouse of information throughout the life course, mo- Because collagen is formed from amino acids, stable iso- bility, ancient weaning practices, and diet change during tope signatures from collagen primarily reflect protein an individual’s lifespan can be studied (Sealy et al., sources in the diet (Ambrose and Norr, 1993; Tieszen 1995; Balasse et al., 1999; Bell et al., 2001; Richards and Fagre, 1993). For a more detailed discussion of sta- et al., 2002; Fuller et al., 2003; Herrscher, 2003; Linder- ble isotopes in anthropology, a number of excellent holm et al., 2008a; Salamon et al., 2008; Jørkov et al., reviews are available (Katzenberg, 2000; Ambrose and 2009; Berger et al., 2010; Nitsch et al., 2011). Thanks to Krigbaum, 2003; Schoeninger, 2011). the ability to study childhood diet using the skeletal remains of adults (i.e., children who lived, rather than Mineralized tissue remodeling children who died), stable isotope biochemistry is uniquely well-suited to circumvent the osteological para- Stable isotope ratios of foods are preserved in the bone dox and explore life history of past populations. When chemistry of consumers, yet bone is not a static reservoir applied to a period of intense cultural change and social of these chemical signatures. Rather, bone is a dynamic pressure such as the European medieval period, bone tissue that changes during life in response to stress. As biochemistry can provide insights on social inequality a result of bone modeling (the formation of new bone in and food access throughout an individual’s lifetime. response to mechanical loading) and remodeling (the Because of its well-documented social stratification maintenance replacement and repair of bone), much of and excellent preservation, the medieval population from the bone collagen formed in younger years disappears Trino Vercellese, Italy, is ideally suited for this kind of with time, and the skeleton is composed of increasingly investigation. Few medieval samples exist that permit more ‘‘new’’ bone. Different skeletal elements exhibit dif- such clear identification of different status groups, nor ferent remodeling rates in relation to their mechanical which are accompanied by such rich archaeological infor- loading and microarchitecture (Stout and Lueck, 1995), mation supporting proper interpretation of the results, indicating that remodeling rates are higher among nor for which such extensive paleopathological and ‘‘active’’ individuals. Unlike bone, enamel and dentine do growth data are available. To interpret stable isotope not remodel during life, and isotopic signatures in these results, it is fundamental to have as much biocultural in- substrates record information about childhood diet (with formation on the population as possible. Trino Vercellese the exceptions of secondary and tertiary dentine). Com- is a rare site where such data—including palynology, paring stable isotope signatures in dentine and bone archeozoology, archeobotany, demography, pathology, and from a single adult has thus the potential to reveal diet material culture analyses—are available. change or residential relocation during an individual’s lifespan (Balasse et al., 1999; Fuller et al., 2003; Sala- mon et al., 2008; Berger et al., 2010; Chenery et al., 2010). GOALS The first goal of this research is to reconstruct diet of MATERIALS a socioeconomically diverse medieval population at Trino The medieval population from Trino Vercellese Vercellese to shed light on the complex dynamics of food access in medieval society. Most previous stable isotope This study includes skeletal materials from the medie- research in Italy focuses on the Roman period (Prowse val site of Trino Vercellese (VC), Italy. This site complex et al., 2004, 2005, 2007; Craig et al., 2009; Rutgers is located in the region and consists of a forti- et al., 2009; Crowe et al., 2010; Nitsch and Hedges, fied settlement that developed around San Michele’s 2010) with a limited number of reports from the Bronze church and its cemetery (Fig. 1). Starting around the Age, Neolithic, and medieval periods (Le Bras-Goude 8th c. and until the 13th c. AD, due to obligatory inter- et al., 2006; Salamon et al., 2008; Tafuri et al., 2009; ment in its grounds of all individuals belonging to the Nitsch and Hedges, 2010). Beyond reconstructing diets, parish, the church of San Michele became the funerary we also investigate sex- and status-based differences in epicenter of the area surrounding Trino Vercellese. Exca- diet and diet change between childhood and adulthood vations carried out by the University of Torino between using dentine and bone collagen in tandem. Where rele- 1980 and 1994 led to the recovery of a total of 749 buri- vant, we also consider the relationship between diet and als located inside and outside the church (Mancini, paleopathology, drawing from previous research at Trino 1999). The majority of the burials (n 5 688) dated to the Vercellese (Celoria, 1999; Girotti and Garetto, 1999; medieval period (8th–13th c.), while a minority of the Vercellotti et al., 2011). burials recovered inside the church was postmedieval

American Journal of Physical Anthropology DIET AND LIFE HISTORY AT TRINO VERCELLESE 591 local consumption but also for regional trade. Faunal remains at the site indicated that the most abundant domestic species were swine, cattle, and sheep/goats, fol- lowed by horses and domestic fowl (Ferro, 1999). Based on butcher marks and age-at-death profiles, it was deter- mined that swine and cattle were mostly bred for meat and hides, while sheep and goat were primarily a source of wool and milk (Aimar et al., 1999). The consumption of wild ungulates, primarily red deer (Cervus elaphus) and roe deer (Capreouls capreouls) also was common- place, because these animals were sought for their ant- lers and hides. Fish remains are also documented, although aquatic resources were probably not a major component of the everyday diet. Anthropological analyses suggested that the medieval population from Trino Vercellese experienced relatively good living conditions, without major growth disruption and with an overall varied and rich diet (Celoria, 1999; Girotti and Garetto, 1999; Porro et al., 1999). This not- withstanding, sex and status-based differences in so- matic growth and oral health were observed (Porro et al., 1999; Vercellotti et al., 2011), suggesting underlying sociocultural differences in terms of access to resources and exposure to disease. We sampled collagen from teeth and bones of 30 (20 males; 10 females) adult individuals from medieval Trino Vercellese, for a total of 60 samples for isotopic analyses. Because accurate knowledge of age, sex, and status is Fig. 1. Map of Italy showing the location of Trino Vercellese absolutely essential for exploring complex food dynamics (black solid circle) and the planimetry of the site [modified from in medieval society, and thus for accomplishing our Mancini (1999)]. In the map of the site, the dark gray area rep- goals, we adopted strict selective criteria for identifying resents the church of San Michele, the light gray area repre- suitable individuals to be included in this study. Specifi- sents the cemetery outside the church, and the black line indi- cally, selection was based on (1) skeletal completeness, cates the fortification walls. defined in terms of preservation of at least 70% of the skeleton—based on this criterion, 52 skeletons from well-defined, individual burials for which accurate infor- (14th c.). Among the medieval burials, 585 (85%) were mation on sex, age, and status were available for the adults and 103 (15%) were juveniles. Based on the demo- analysis; (2) availability of both ribs and second molars graphics of the adult cemetery, the skeletal collection is (M2)—based on this criterion, a total of 37 skeletons considered to be a representative sample of Trino Ver- were suitable for sampling; (3) minimal dental wear—so cellese’s adult inhabitants during the medieval period as to avoid issues related to sampling of secondary and (Mancini, 1999; Vercellotti et al., 2011). Furthermore, it tertiary dentine, all individuals whose M2s were has been advanced that the skeletal sample recovered extremely worn or affected by carious lesions could not from Trino is representative of the socioeconomic varia- be included, reducing our sample to 32. Of the 32 skele- tion expressed by the population during the medieval pe- tons meeting all selective criteria, we chose 30 individu- riod. Several lines of evidence suggest that groups of dif- als so as to equally represent the different status groups. ferent social status were buried in different areas of San With these strict selection criteria, data from all 30 indi- Michele’s cemetery, supported by differences in (1) burial viduals can contribute to broader parallel investigations location, (2) burial typology, and (3) grave good typology into not just diet, but also health and nutrition at Trino [for a detailed review of this evidence, see Vercellotti Vercellese. et al. (2011)]. To represent both sexes and status groups equally, a Extensive archaeological, archaeozoological, palaeobo- similar number of high and low-status burials were tanical, and anthropological studies have been carried sampled, based on availability of suitable individuals: out on the site, allowing for an extremely detailed recon- (1) high status (4 females, 10 males) and (2) low status struction of environmental conditions and lifestyle of the (6 females, 10 males). Status groups were determined medieval population inhabiting Trino Vercellese (Accorsi based on burial location and typology, and the presence et al., 1999; Aimar et al., 1999; Caramiello et al., 1999; of grave goods, as described in detail elsewhere (Vercel- Celoria, 1999; Ferro, 1999; Girotti and Garetto, 1999; lotti et al., 2011). Sex and age-at-death of all individu- Mancini, 1999; Porro et al., 1999). Overall, this research als were estimated from sexually dimorphic features of indicates that throughout the medieval period, the area the pelvis (Phenice, 1969; Buikstra and Ubelaker, surrounding Trino Vercellese was characterized by hard- 1994), morphological alterations of the os pubis’ articu- wood forests, the extension of which was progressively lar surface (Brooks and Suchey, 1990), and the ilium reduced by anthropic deforestation in favor of pastures auricular surface (Lovejoy et al., 1985). Age-at-death and crops (Caramiello et al., 1999). Primary crops were of all individuals for all individuals fall between 20 represented by cereals, legumes, and aromatic plants and 55 years. The age structures of the four (Umbelliferae; Accorsi et al., 1999). The local economy subsamples show no significant difference (Kruskal–Wallis, was centered on livestock breeding, destined not only for P 5 0.51).

American Journal of Physical Anthropology 592 L.J. REITSEMA AND G. VERCELLOTTI

TABLE 1. Comparative fauna used for a dietary baseline Sample d15N[%] d13C[%] Reference Southeastern France; medieval 4.9 6 1.0 220.8 6 0.6 Eleven animals (cow, pig, sheep, goat, chicken; Herrscher et al., 2001) Northern Italy; Bronze Age Olmo di Nogara 6.4 6 1.0 216.6 6 1.2 Three animals (cow, pig, goat; Tafuri et al., 2009) Mereto 4.5 220.5 One cow (Tafuri et al., 2009)

METHODS obtained, and %coll can only be estimated at less than Sample collection and preliminary treatment 3.2%. Dried collagen was homogenized using an agate mor- To investigate diet change over the lifetime at Trino tar/pestle. Between 0.600 and 0.700 mg of powdered den- Vercellese, we examined stable carbon and nitrogen iso- tine/bone collagen for each sample was analyzed on a tope values from cortical bone of a rib along with a tooth Costech Elemental Analyzer coupled to a Finnigan Delta for each individual included in the study. Subsequent IV Plus stable isotope ratio mass spectrometer under identification numbers for individuals are given as the continuous flow using a CONFLO III interface in the burial number, tooth or bone initial, and sex initial (e.g., Stable Isotope Biogeochemistry Laboratory at The Ohio S-133-T-M, a tooth from the male individual #133). We State University. Approximately 10% of all samples were 13 chose to sample M2, which form between 3.7 and 6.7 run in duplicate. Stable carbon (d C 5 permil deviation 13 12 years of age on average (Moorrees et al., 1963), to pro- of the ratio of C: C relative to the Vienna Peedee vide information on diet during childhood. Medieval Belemnite Limestone standard) and stable nitrogen 15 15 14 records indicate a recommended weaning age of 2 years (d N 5 permil deviation of N: N relative to AIR) iso- (Shahar, 1990), which has been supported by stable tope measurements were made where the average stand- isotope studies (Richards et al., 2002); thus, the isotopic ard deviation of repeated measurements of the USGS-24, 13 signature of M2 is expected to reflect postweaning child- IAEA-N1, and IAEA-N2 standards were 0.05% for d C 15 hood diets at Trino Vercellese. The majority of teeth and 0.13% for d N. Stable isotope ratios are expressed sampled were still in situ in the alveolar bone. Mandibu- as a permil (%) ratio of one of an element’s isotopes to lar M2 were sampled preferentially, and from the left another in relation to a standard of known abundance 13 15 side, when possible; if absent, maxillary M2 were (Vienna Pee Dee Belemnite for d C and AIR for d N). 13 15 sampled. Casts were made of all teeth before preparing Both carbon (d C) and nitrogen (d N) stable isotope them for isotopic analysis. ratios are reported according to the equation [d 5 3 Each tooth was prepared by resecting the root at the (Rsample – Rstandard)/Rstandard 1,000]. cementoenamel junction using a Piezosurgery1 3 device Animal bones were not made available and conse- (Piezosurgery Incorporated, Columbus, OH). This tech- quently were not directly sampled for this study. nology uses modulated ultrasonic frequencies to cut min- Instead, a faunal baseline is estimated using two previ- eralized tissues with extreme precision while respecting ously published samples (Table 1). The first sample com- soft tissues. This technology was selected, because it prises 11 animals (cow, pig, sheep, goat, chicken) from allows resecting and scraping samples with extreme pre- an inland medieval site from Grenoble, Ise`re (France) cision and safety for the operator. Upon removal of the located  250 km due west of Trino (Herrscher et al., root, crowns were sectioned in four pieces, and any sec- 2001). The second comprises four animals (cow, pig, ondary dentine lining the pulp cavity (easily recognized goat) from two Bronze Age sites in Northern Italy due to its darker color) was entirely removed using an (Tafuri et al., 2009). ultrasonically operated scraper. No special effort was made to remove the enamel, although it chipped in some Statistical analyses cases and was then removed. To evaluate and compare stable isotope signatures among sex and status subsamples, we used the Mann– Preparation of samples for isotopic analysis Whitney or Kruskal–Wallis tests, depending on the num- ber of subsamples being compared. These nonparametric Between 0.30 and 1.00 g of whole tooth and rib pieces tests are preferable to the parametric ANOVA, because were demineralized in 1% HCl, rinsed, and soaked in they are applicable even when the distribution of the 0.125 M NaOH to remove humic contaminants (Richards data departs from normality (Zar, 1999). Statistical and Hedges, 1999; Garvie-Lok, 2001). In most cases, analyses were performed with MYSTAT Software. In the structural integrity of bone and dentine pieces was pre- following sections, P-values are considered statistically served, but, in two cases (S-424-B-M and S-535-B-F), the significant if less than or equal to 0.05. model had crumbled into debris during chemical leach- ing. Bone and dentine pieces were rinsed, dissolved over- RESULTS night in dilute HCl (pH 5 3) in a 908C oven, centrifuged, Sample quality and freeze-dried. Weights of dried bone collagen were obtained to estimate collagen concentration in bone To evaluate diagenesis, five measurements are consid- (%coll). In one case (S-41-B-M), a small amount of sam- ered (1) carbon content in collagen (%C), (2) overall ple leaked from the vial during lyophilization, so its nitrogen content in collagen (%N), (3) atomic carbon to actual %coll is somewhat greater than the 3.3% meas- nitrogen (C:N) ratios calculated as C:N 5 [(%C/12)/(%N/ ured. In another case (S-68-B-M), the amount of collagen 14)], (4) collagen content in bone (%coll; not measured in remaining was less than the tare of the scale used dentinal collagen due to variable amounts of enamel (\0.01 g) indicating a problem with the initial weight affecting initial weights), and (5) preservation of a colla-

American Journal of Physical Anthropology DIET AND LIFE HISTORY AT TRINO VERCELLESE 593

TABLE 2. Stable isotope medians and ranges for four demographic subgroups d15N(%) d13C(%) Dentine Bone Dentine Bone Subgroup Median Range Median Range Median Range Median Range Males (n 5 19) 9.5 8.0–12.0 8.8 8.1–11.8 219.4 220.0 to 217.6 218.9 219.9 to 217.4 High-status (n 5 10) 9.4 8.9–10.4 9.2 8.6–10.3 219.5 219.6 to 218.2 219.3 219.9 to 218.5 Low-status (n 5 9) 9.6 8.0–12.0 8.4 8.1–11.8 219.1 220.0 to 217.6 218.5 219.3 to 217.4 Females (n 5 9) 9.2 6.7–10.3 9.2 9.0–10.5 219.4 220.1 to 218.5 219.5 219.9 to 219.1 High-status (n 5 4) 9.2 9.0–9.2 9.3 9.1–10.5 219.3 219.9 to 218.5 219.8 219.9 to 219.4 Low-status (n 5 5) 9.5 6.7–10.3 9.1 9.0–9.4 219.4 220.1 to 218.6 219.3 219.8 to 219.1

15 gen model of the original bone piece(s) following the HCl (d Nbone) ranged from 8.1 to 11.8% (mean 5 9.2 6 13 13 and NaOH components of sample preparation. Details 0.8%). d C values in dentine (d Cdentine) ranged from – on the rationale for acceptable ranges for these measure- 20.1 to –17.6% (mean 5 –19.2 6 0.7%) and in bone 13 ments are described elsewhere in greater detail (d Cbone) ranged from –19.9 to –17.4% (mean 5 –19.1 6 (Ambrose, 1990; Garvie-Lok, 2001). With the exception 0.7%). A complete data set is presented in Table S1. of two bone samples, all bone and tooth samples show Summary data, including medians and dispersion little or no signs of diagenetic alterations. Sample S-424- (range) for each of the four subgroups in accordance B-M did not yield an intact collagen model, exhibited low with the nonparametric statistics used, are presented in %C (3.7%), low %N (0.7%), and a very high C:N value of Table 2. 6.1. Sample S-535-B-F also did not yield an intact colla- gen model, exhibited a relatively high C:N value of 3.7, Dentine. Dentine stable isotope values are presented in and yielded a %coll value of just 1.00%. Sample S-535-B- Figure 2 divided into sex and status groups. For the F yielded only the %coll value to fall below a threshold pair-wise comparisons among subgroups using the level of 2% described by Ambrose (1990). These two bone Mann–Whitney U test statistic, Table 3 displays P-val- samples are excluded from further analyses, although ues, with statistically significant values (P = 0.05) the dentine values from these individuals are included appearing in bold. The d15N ratios did not differ where possible. dentine significantly by sex (males: 9.5 6 0.9%; females: 9.1 6 For the 58 remaining bone and tooth samples, C:N 1.0%; P 5 0.481). The d13C ratios also did not dif- ratios range from 3.2 to 3.5. Collagen concentration in dentine fer significantly by sex (males: –19.1 6 0.7%; females: – bone (%coll; only measured for bone) falls between 2.6 19.3 6 0.6%; P 5 0.367). and 22.5%. For bones, %N values fall between 2.8 and Neither d13C nor d15N ratios differ signifi- 14.8%, and %C values fall between 8.4 and 41.8%. For dentine dentine cantly by status, when males and females are grouped teeth, elemental concentrations are generally higher and together in status groups. The mean d15N value of show more restricted ranges: tooth %N ranges from 13.1 dentine high-status individuals is 9.4 6 0.4% compared to 9.3 6 to 15.9% and tooth %C ranges from 29.1 to 43.7%. Two 1.2% for low-status individuals (P 5 0.724). Despite no samples have %N and %C values below the ‘‘well-pre- statistically significant differences, all the lowest dentine served’’ values of 4.8 and 13.0%, respectively, for prehis- d15N values are exhibited by low-status individu- toric bones reported by Ambrose (1990; S-134-B-M and S- dentine als. Also, dentine of high-status individuals shows a 57A-B-M). However, both exhibit acceptable C:N and much narrower d15N range (9.0–10.4%) than dentine of %coll values and yielded intact collagen models and are low-status individuals (6.7–12.0%). The mean d13C not excluded from subsequent analyses. There are no cor- dentine value of high-status individuals is –19.2 6 0.5% com- relations between any of the collagen quality indicators pared to –19.1 6 0.8% for low-status individuals (P 5 (except %C and %N, as expected), indicating that diage- 0.724). The range of d13C values is also narrower netic alteration of stable isotope signatures in the sample dentine for high-status individuals than for low-status individu- is minimal. Collagen quality indicators for all individu- als, although not to the same extent as for d15N . als, including the two whose bone data were excluded on dentine When status groups are further divided by sex, there the basis of poor collagen quality (demarcated), are are no significant differences between dentine stable iso- reported in Table S1 along with stable isotope data. tope values of high- versus low-status females or Seven duplicates (four bones and three teeth) of the between high- versus low-status men. Lack of statistical same collagen powder were analyzed to estimate a margin significance could be due to the small female subsample of analytical error. Within this margin of error, differences sizes, although the overall variation among females is in stable isotope signatures may not be reliably inter- also small and, excepting one high-d15N outlier (S-48A-F; preted as biologically meaningful. For d15N, the maximum d15N 5 10.5%), all female values are within the maxi- difference between two duplicates was 0.7%, and the mum margin of error for duplicate analyses. mean difference was 0.3%. For d13C, the maximum differ- ence between two duplicates was 0.1%, and the mean dif- ference was less than 0.1%. Both the mean and the maxi- Bone. Bone stable isotope values are presented in Fig- mum differences between duplicates to represent margins ure 3 and divided into sex and status groups. The d15N 15 of error are shown in figures in the following sections. values of bones (d Nbone) did not differ significantly by 15 sex. The mean d Nbone value for males was 9.1 6 0.9% and for females was 9.3 6 0.5% (P 5 0.134). However, Stable isotope results 13 d Cbone did differ significantly by sex, with males exhib- 15 15 Overall, d N values in dentine (d Ndentine) ranged iting on average higher values: –18.8 6 0.7% compared from 6.7 to 12.0% (mean 5 9.4 6 0.9%) and in bone to –19.6 6 0.3% for females (P 5 0.005).

American Journal of Physical Anthropology 594 L.J. REITSEMA AND G. VERCELLOTTI When considering each status group separately, sex- based differences in bone stable isotope ratios are most pronounced among low-status individuals. In both status 13 groups, females exhibit lower d Cbone values than do males, but the difference is most pronounced among low- status individuals (low-status males versus females, P 5 0.009, compared to P 5 0.066 for high-status male ver- sus female individuals). With one exception (S-408-B-M; 13 13 low d C), there is no overlap between d Cbone values of males and females among the low-status individuals, whereas there is considerable overlap between high-sta- 13 tus males’ and females’ d Cbone values (Fig. 3). Comparing Figures 2 and 3, there appears to be evi- dence that childhood diets were more variable than adult diets, particularly for individuals of low status. The 15 Fig. 2. Stable isotope data from dentine of M2, representing d Ndentine values show greater dispersion than do 15 childhood diet, of the four subgroups investigated (high- and d Nbone values, although this is mostly the effect of a low-status men and women). 15 15 single low- Ndentine outlier (S-528-F; d N 5 6.7%). When this individual is removed, the d15N ranges of the overall sample are nearly the same (dentine: 8.0–12.0%; bone: 8.1–11.8%). The d13C ranges of dentine and bone are also nearly the same (dentine: –20.1 to –17.6%; bone: –19.9 to –17.4%). Dentine-bone spacing relationships. Figures 4 and 5 display the difference between teeth and bones from the same individual, as calculated by subtracting the bone stable isotope value from the dentine value. On average, 15 15 15 the d Ndentine-d Nbone (d N) absolute values of low-sta- tus individuals are greater and more variable than d15N absolute values of high-status individuals. The mean d15N absolute value of low-status individuals is 0.8 6 0.9%, whereas the mean absolute value of high-status individuals is just 0.4 6 0.4% (P 5 0.346). The same is 13 13 13 13 true for d C: the mean d Cdentine-d Cbone (d C) differ- Fig. 3. Stable isotope data from bone collagen (rib), repre- ence of high-status individuals is 0.5 6 0.4% and of low- senting adult diet, of the four subgroups investigated (high- and status individuals is 0.7 6 0.7% (P 5 0.535). The d dif- low-status men and women). ferences between high- and low-status groups are the product of a few high-d outliers within the low-status subgroup and are not statistically significant. There are differences in both d15N and d13C of bones The direction of these dentine-bone stable isotope of high- and low-status individuals (males and females shifts throughout life is depicted in Figures 4 and 5. In 15 grouped together). The mean d Nbone value for high-sta- each of these figures, dashed lines above and below the tus individuals is 9.4 6 0.6%, compared to 9.0 6 1.0% main axis represent the mean d difference obtained for 13 for low-status individuals (P 5 0.060). The mean d Cbone all duplicate analyses, and solid lines above and below value of high-status individuals is –19.4 6 0.4% com- the main axis represent the maximum difference pared to –18.8 6 0.8% for low-status individuals (P 5 obtained for all duplicate analyses. 0.024). Bones of younger individuals should retain relatively When considering each sex separately, bones of high- more collagen from childhood than bones of older indi- and low-status females do not differ significantly in ei- viduals, whose bones are more highly remodeled. How- ther d15N(P 5 0.133) or d13C(P 5 0.221), although the ever, there is no correlation between age and either d15N female subsamples are admittedly small. However, high- (R2 5 0.0443) or d13C(R2 5 0.0585), suggesting that 15 status males exhibit significantly higher d Nbone (P 5 bone had remodeled enough even by the youngest ages 13 0.034) and significantly lower d Cbone (P 5 0.009) values of 20 to record an adult diet that differed from childhood than do low-status males. diet, as expected (Hedges et al., 2007).

TABLE 3. Results of pair-wise comparisons using the Mann–Whitney U test statistic Dentine Bone Sex and status comparisons d15N d13C d15N d13C Males versus females—status groups combined P 5 0.481 P 5 0.367 P 5 0.134 P 5 0.005 High status—males versus females P 5 0.322 P 5 0.671 P 5 0.396 P 5 0.066 Low status—males versus females P 5 0.588 P 5 0.193 P 5 0.096 P 5 0.009 High status versus low status—sexes combined P 5 0.724 P 5 0.724 P 5 0.060 P 5 0.024 High status versus low status—males P 5 0.970 P 5 0.496 P 5 0.034 P 5 0.009 High status versus low status—females P 5 0.394 P 5 0.522 P 5 0.221 P 5 0.133

Highlighted P values are significant at the 0.05 level

American Journal of Physical Anthropology DIET AND LIFE HISTORY AT TRINO VERCELLESE 595 evoke freshwater fish consumption for individuals in 3rd–5th c. Roman catacombs whose d15N values are higher than 11.5% and whose d13C ratios are lower than –19.5%. None of the individuals in the present study ex- hibit values such as these. In the recent years, there has been a growing appreciation for the complexity of aquatic isotope environments and the fact that fish from both marine and freshwaters may exhibit isotope values similar to those of terrestrial animals (Katzenberg et al., 2010; Bourbou et al., 2011). Although the evidence is not strong, in the present study, the wide range in d15N values within a relatively narrow range of d13C values (–19.5 to –20.1%) could suggest that people were eating small amounts of fish with d15N values of  6–9% and Fig. 4. Differences between d15N values obtained from teeth d13C values similar to those of terrestrial animals, such as and bones of each individual, divided into four subgroups (high- have been reported previously for the Mediterranean and low-status men and women). The solid line represents the region and some freshwaters (Grupe et al., 1999; Prowse d15 % maximum difference between two N duplicates (0.7 ), and et al., 2004; Keenleyside et al., 2006; Reitsema et al., the dashed line represents the mean difference between two 2010; Bourbou et al., 2011). Detecting fish such as these in d15N duplicates (0.3%). Negative d15N values represent an increase in d15N between childhood and adulthood, whereas pos- human diet remains a challenge in stable isotope studies itive d15N values represent a decrease in d15N. (e.g., Privat et al., 2002; Mu¨ ldner and Richards, 2005). The paucity of fish in diet in the present study is noteworthy. At this time in Europe, Christian fasting regulations provided an impetus to replace protein from terrestrial animals with fish (Woolgar, 2000). Fish are documented to increase in diet with Christianity else- where in Italy and Europe and may be more common in diets of the clergy than of the general populace (Polet and Katzenberg, 2003; Barrett and Richards, 2004; Sala- mon et al., 2008; Mu¨ ldner et al., 2009; Rutgers et al., 2009). Even among bones of high-status individuals at Trino Vercellese buried in the church, there is no com- pelling evidence for fish consumption. Although fish were sometimes considered a luxury in medieval Europe, in other cases, they were regarded as an undesirable food or a marker of low status (Ervynck et al., 2003; Van Neer and Ervynck, 2004). In this regard, it is interesting to notice the cultural belief in medieval Northern Italy Fig. 5. Differences between d13C values obtained from teeth and bones of each individual, divided into four subgroups (high- that fish from stagnant bodies of water were to be and low-status men and women). The solid line represents the avoided, as they were considered unhealthy (Nada maximum difference between two d13C duplicates (0.12%), and Patrone, 1981). In an area containing many smaller slow the dashed line represents the mean difference between two moving fluvial branches and ditches, yet lacking major d13C duplicates (0.05%). Negative d13C values represent an streams and rivers such as medieval Trino Vercellese, increase in d13C between childhood and adulthood, whereas pos- fish would therefore have been marginally exploited as a 13 13 itive d C values represent a decrease in d C. food source. This seems to be corroborated by the paucity of fish remains from the site (Ferro, 1999). The only fish remains recovered in Trino belong to the northern pike DISCUSSION (Esox lucius), a fish locally found in slow-moving riverine Diet environment, the consumption of which was likely not advisable by Piedmont medieval cultural dietary beliefs. Stable isotope data support an interpretation of a ter- In light of this, Christian fasts may not have been rig- restrial-based diet including both plant and animal pro- idly observed. Based on the complex medieval ideology tein. Compared to the faunal baseline, human d15N val- centered on meat consumption as a status symbol (Mon- ues are generally consistent with one trophic level of tanari, 1988), it is plausible that high-status individuals enrichment. The human data are most similar to results did not abstain from consuming animal meat despite reported by Richards et al. (1998) for wood-coffin burials Church directives. According to Montanari (1988), even in the Roman period cemetery at Poundbury, UK (n 5 members of the clergy—as high profile elites them- 26) (Fig. 6), which were interpreted as indicating a ter- selves—would have had social and political reasons not restrial C3 diet. There is little evidence for fish consump- to observe Christian fasts. It is also possible that while tion. All individuals exhibit d15N values lower than meat from terrestrial animals was eschewed during 12.5%, a terrestrial–marine cutoff used by Salamon et fasts, it was replaced with protein from dairy products al. (2008) for a medieval sample from Rome. This is per- (milk, cheese, eggs), the stable isotope signatures of haps not surprising, as Trino Vercellese is an inland site. which do not differ appreciably from meat (Steele and Freshwater fish may have been reasonably accessed by Daniel, 1978; Garvie-Lok, 2001). the population in local streams, but the archaeological The wide range in d15N values could also reflect con- and stable isotope evidence for freshwater fish consump- sumption of omnivore protein or protein from suckling tion is not strong (Ferro, 1999). Rutgers et al. (2009) animals, both of which may exhibit d15N values as high

American Journal of Physical Anthropology 596 L.J. REITSEMA AND G. VERCELLOTTI

Fig. 6. Individual bone samples from Trino Vercellese are shown in comparison with other European populations, which are dis- played as mean 6 1 standard deviation. Individuals from Trino Vercellese overlap with populations consuming terrestrial-based diets. Some individuals overlap with a Northern Italian population believed to consume the C4 plant millet. One individual with a high-d15N value plots with a population-consuming marine fish; this individual likely moved to Trino Vercellese from the coast, as discussed in the text.

TABLE 4. Cereal species present in Piedmont and at Trino Vercellese during the medieval period (after Nada Patrone, 1981 and Accorsi et al., 1999) Photosynthetic Archaeologically present Cereal Scientific name Grain type pathway at Trino Vercellese Rye Secale cereale Large C3 Present Wheat Triticum sp. Large C3 Present Barley Hordeum sp. Large C3 Present Proso millet Panicum miliaceum Small C4 Present Foxtail millet Panicum italicum (Setaria italica) Small C4 Absent Sorghum Sorghum sp. Small C4 Absent as 91% due to their elevated trophic position (Mu¨ ldner mon in the provinces of and Vercelli (hence at and Richards, 2007). Historical accounts of dietary pref- Trino Vercellese) than in other areas of the region, erences and culturally mandated dietary behaviors in likely due to their geographical proximity to Lombardia, Piedmont in the medieval period support this explana- where millet was much more common (Nada Patrone, tion (Nada Patrone, 1981). Indeed, according to the cul- 1981). Analyses of the paleobotanical evidence from ture of the time, it was preferable for sociocultural elites Trino Vercellese indicate that a number of different to consume the meat of young animals (such as lambs, cereals were cultivated at this site (Accorsi et al., 1999; kids, calves, and young fowl). The presence of large num- Caramiello et al., 1999; Nisbet, 1999), including millet bers of subadult animals bearing butcher marks at Trino (Table 4). Vercellese (Ferro, 1999) suggests that such dietary cus- Based on this evidence, it is reasonable to assume that toms were applied here as well. Birds, too, may exhibit some individuals at Trino Vercellese may have consumed 15 13 variable d N values due to their diets and habitats millet, a C4 cereal that has a d C value of  –10 to – (Grupe et al., 1999). Human consumption of either 12% (modern values) (McGovern et al., 2004). Stable iso- 13 domestic or wild fowl could contribute to the observed tope values of the high- Cbone group overlap with data variable d15N values accompanying restricted d13C val- reported by Tafuri et al. (2009) for millet-consuming ues at Trino Vercellese. individuals from Bronze Age Northern Italy (the site of Several individuals from Trino Vercellese exhibit Sedegliano; Fig. 6). Without a local faunal baseline, it is 13 13 d Cbone values higher than –18.0%. These individuals not yet possible to discern whether the high- C signal 15 also exhibit relatively low d Nbone values (Fig. 3). in human collagen is from consumption of millet or of Because fish consumption is an unlikely explanation in animals foddered on millet. Tafuri et al. (2009) observed 13 this case, C enrichment is probably due to a C4 plant a much stronger ‘‘millet signal’’ in bones of Bronze Age in the diet. The presence of C4 cereals such as millet individuals from Olmo di Nogara in Northern Italy, (including Panicum miliaceum and Panicum italicum, where the faunal baseline clearly demonstrated that mil- also known as Setaria italica), and sorghum in the Pied- let was used as a fodder. Because collagen preferentially mont during the medieval period is well documented. In reflects protein sources in diet, it may be the case that particular, the cultivation of millet was much more com- the humans studied by Tafuri et al. (2009) exhibit

American Journal of Physical Anthropology DIET AND LIFE HISTORY AT TRINO VERCELLESE 597 greater enrichment in 13C, because the signal came statistically significant female status differences is con- from protein, whereas in the present study, the high-13C sistent with previously reported evidence from this popu- foods are underrepresented in collagen because of lation from indicators of developmental stress (linear direct consumption of millet and not consumption of enamel hypoplasia and adult stature) and dental pathol- millet-foddered animals. Also, it would seem that if ani- ogy (tooth loss and dental caries) showing pronounced mals at Trino Vercellese were fed millet, the ‘‘millet sig- differences in developmental stress and poor oral condi- nal’’ would be even stronger among individuals who con- tions between high- and low-status males, but almost no sumed more animal protein; that is, a positive correla- differences between high- and low-status females (Girotti tion may be expected between d13C and d15N, which is and Garetto, 1999). Evidently, low-status females con- not observed. Comparative analysis of d13C from bone sumed diets more similar to both high-status females apatite, which more equally reflects all dietary macronu- and high-status males, including more animal protein trients (Krueger and Sullivan, 1984; Ambrose and Norr, and less millet than diets of low-status males. 1993; Tieszen and Fagre, 1993), would help clarify the Another notable difference between the diets of high- question of millet in diet, as has been done elsewhere and low-status adults is that bone d values of low-status (Reitsema et al., 2010). individuals show greater variation, whereas values of high-status adults are more restricted (Fig. 3). A similar situation is reported for Roman period UK (Richards et Status- and sex-based differences al., 1998), where high variability among low-status indi- 15 viduals was attributed to diverse diets characteristic of A narrower range of high-status d Ndentine values sug- gests that high-status children had more consistent diets life in an urban center. Lower isotopic variation for high- within their class, whereas the diets of low-status chil- status individuals than for low-status individuals is also dren were unselective. Status-based differences in reported for early medieval Bavaria (Czermak et al., females’ diets are not apparent; however, sample sizes 2006), where it is attributed to elites having more con- may be too small to detect significant differences. High- sistent access to animal protein. Similarly, at Trino, vari- status adult males consumed more animal protein than ability in low-status diets is probably due to individual did low-status adult males. High-status males likely fortunes and misfortunes, and the ever-present uncer- include members of the clergy and local nobles who in tainty of the access to resources experienced by low-sta- life were of higher socioeconomic status (Negro Ponzi tus individuals in the medieval period. d15 Mancini, 1999; Vercellotti et al., 2011). During the medi- Interestingly, the highest bone and dentine N are eval period throughout Europe, protein from terrestrial from a single low-status individual (S-408-M). His stable animals was generally more expensive than other foods isotope values are very similar to those reported for indi- and thus restricted to the sociopolitical elite (Adamson, viduals consuming marine fish at the coastal site of Velia 2004; Dembinska, 1999; Dyer, 1994). At Trino Vercellese in Southern Italy (1st–2nd c. A.D.; Craig et al., 2009). as elsewhere (Kjellstro¨m et al., 2009; Linderholm et al., We propose that this individual may have spent most of 2008b; Polet and Katzenberg, 2003; Schutkowski et al., his life at a coastal site, consuming more marine protein 1999), these status-based medieval dietary differences than the rest of the studied individuals. This individual are visible in bone chemistry. was buried in an area (Sagrato Nord) that was recog- Another status-based difference is the greater millet nized by the archaeologists as distinct from the rest of intake by low-status adult males. The four individuals the cemetery. Specifically, this area had 24 individual showing the strongest evidence for millet consumption burials organized in clusters and disposed in parallel (see Fig. 3) are all males buried outside the church, lines. Additionally, whereas sex ratio in the rest of the whose diets are also relatively low in animal protein. cemetery is about 50:50; in this separate area, skeletons This is consistent with historical reports of millet as a are 80% male. Although no specific interpretation of the low-status food (Nada Patrone, 1981; Spurr, 1983; Adam- formation of this separate area was advanced, it was son, 2004). An historical examination of cereal produc- observed that, like anomalous burial clusters in other tion and consumption in Piedmont during the medieval medieval cemeteries, the area may have formed during period reveals the presence of both large grain cereals the transition from family to parish cemetery (or vice (rye, wheat, barley) and small grain cereals (millet, sor- versa; Mancini, 1999). ghum). It is interesting to note that, according to land use contracts, small grain cereals were subject to lower Life history fees than large grain cereals, which were considered more valuable. This was the expression of a food culture Stable isotope data support an interpretation of simi- that reserved large grain cereals—in particular Triti- lar diets between status groups and the sexes during cum—to the higher class, while consumption of small childhood, with increasing disparities into adulthood, as grain cereals was limited to the populace (Nada Patrone, estimated from bones. In general, the disparity consists 1981). As a consequence, landowners had a diet that dif- of reduced access to animal protein for low-status males fered from that of their workers also in terms of choices throughout life (Fig. 4). Another change in diets of low- between food types. Although the nutritional profiles of status males over time is an increase in millet consump- these different grains differ, the small-grained cereals tion (Fig. 5). Childhood diet may have been more varied cannot be considered less nutritious than the large- than adult diet, although the evidence for this is not grained cereals, such that the status-based difference in strong. their regard should be biologically meaningful (Watt and Compared to what is observed among low-status Merrill, 1975). males, diets of females changed little from childhood to Although differences between high- and low-status adulthood (Figs. 4 and 5). Based on historical evidence males are observed, there are no discernible status-based suggesting that millet was a less-desirable food, children differences in stable isotope ratios of females (Fig. 3). and females may have been treated differently to help The subsamples of females are small, but the lack of guarantee survival. We propose that there may have

American Journal of Physical Anthropology 598 L.J. REITSEMA AND G. VERCELLOTTI been a cultural buffer for females of reproductive age. do not know on what basis diets of males varied, but it Although medieval society was undoubtedly male-domi- stands to reason it could have had something to do with nant, other evidence for female buffering in medieval so- their occupation or ability to earn a living. ciety exists. For example, skeletal indicators of health suggest that females experienced more stable living con- ACKNOWLEDGMENTS ditions in the transition from Late Antiquity to the medi- eval period in Croatia (Sˇ laus, 2008). In medieval Swe- We thank Professor Emma Rabino Massa, Director of den, females consumed more consistent diets than did the Museum of Anthropology and Ethnography in , males, which may also be a case of cultural buffering, and Rosa Boano, Collection Curator, for granting permis- although the authors advance the possibility that sion to study the materials. We are grateful to Donatella females were more stationary than males (Kjellstro¨met Minaldi and Gianluigi Mangiapane for their logistical al., 2009). It is also possible that the relative consistency support during sampling and data collection in Italy. We of medieval females’ stable isotope values in the present thank Andre´a Grottoli and Yohei Matsui of The Ohio study has to do with their regular involvement in day-to- State University Stable Isotope Biogeochemistry Labora- day food preparation tasks. Low-status male laborers, in tory and Douglas E. Crews of The Ohio State University contrast, may have had access to fewer foods—and more Human Biology Laboratory for much analytical support. specific foods (e.g., millet) —simply because of their daily Finally, we thank three reviewers for their feedback on routines outside the home. the manuscript. As with diets of women, there are minimal changes throughout life in diets of high-status individuals (both males and females). This likely reflects a certain degree LITERATURE CITED of certainty or selectivity in high-status diets. Aberth J. 2005. The Black Death: the Great Mortality of 1348- 1350. New York: Bedford/St. Martin’s. CONCLUSIONS Accorsi CA, Mazzanti MB, Forlani L, Caramiello R, Nisbet R. 1999. L’archivio archeobotanico: applicazione dell’archivio al sito arche- Diets of the studied individuals are terrestrial-based, ologico di Trino Vercellese 130 m SLM; 458 120 LAT N 88 180 which could be expected based on geography (Trino Ver- LONG E (Vercelli—, Nord Italia). In: Mancini MM, edi- cellese is an inland site), but is in fact noteworthy con- tor. San Michele di Trino (VC) Dal villaggio romano al castello sidering the various impetuses for fish consumption in medievale, Vol. 2. Firenze: All’insegna del Giglio. p 601–620. medieval Europe (e.g., Christianization, trade connec- Adamson MW. 2004. Food in medieval times. Westport, CT: tions with the coast and feeding aggregated commun- Greenwood. ities). This questions the idea that fish are dietary Aimar A, De Bella A, Ferro AM. 1999. L’analisi tafonomica. In: markers for certain medieval religious and economic Mancini MM, editor. San Michele di Trino (VC) Dal villaggio romano al castello medievale, Vol. 2. Firenze: All’insegna del influences (Mu¨ ldner and Richards, 2005; Barrett et al., Giglio. p 647–652. 2008). Ambrose SH. 1990. Preparation and characterization of bone Differences among the subgroups are relatively small and tooth collagen for isotopic analysis. J Archaeol Sci in childhood as measured from dentine and expand into 17:431–451. adulthood as measured from bone, which may be the Ambrose SH, Krigbaum J. 2003. Bone chemistry and bioarch- effect of cultural buffering of children. The most pro- aeology. J Anthropol Archaeol 22:193–199. nounced changes during the life course are seen among Ambrose SH, Norr L. 1993. Experimental evidence for the rela- low-status males. Low-status adult males differ consider- tionship of the carbon isotope ratios of whole diet and dietary ably from females and high-status males in that they protein to those of bone collagen and carbonate. In: Lambert JB, Grupe G, editors. Prehistoric human bone: archaeology consumed a diet with more millet and less meat, a differ- at the molecular level. Berlin: Springer-Verlag. p 1–37. ence that developed after childhood. Balasse M, Bocherens H, Mariotti A. 1999. Intra-bone variabili- A look at life history of the studied population yields ty of collagen and apatite isotopic composition used as evi- several interesting observations. Evidently, at Trino Ver- dence of a change in diet. J Archaeol Sci 26:593–598. cellese, the diets of high-status individuals and diets of Barrett J, Johnstone C, Harland J, Van Neer W, Ervynck A, females in general are more consistent throughout the Makowiecki D, Heinrich D, Hufthammer AK, Enghoff IB, life course. Perhaps, due to their role in reproduction Amundsen C, et al. 2008. Detecting the medieval cod trade: a and child rearing, females were afforded consistent new method and first results. J Archaeol Sci 35:850–861. access to foods without regard to their socioeconomic sta- Barrett JH, Richards MP. 2004. Identity, gender, religion and tus and without risky fluctuations year-to-year or sea- economy: new isotope and radiocarbon evidence for marine resource intensification in early historic Orkney, Scotland, son-to-season. This consistent food access may have been UK. Eur J Archaeol 7:249–271. in recognition of women’s nutritional needs during their Bell LS, Cox G, Sealy J. 2001. Determining isotopic life history reproductive years, which for medieval females was most trajectories using bone density fractionation and stable iso- of their adult lives, or it could have been the by-product tope measurements: a new approach. Am J Phys Anthropol of providing consistent food access to children, with 116:66–79. whom females may have spent more time day-to-day Berger TE, Peters J, Grupe G. 2010. Life history of a mule (c. than males. This possibility underscores the significant 160 AD) from the Roman fort Biriciana/Weissenburg (Upper potential that bioarchaeology has to nuance interpreta- Bavaria) as revealed by serial stable isotope analysis of dental tions of male-dominant medieval society. tissues. Int J Osteoarchaeol 20:158–171. While high-status males, high-status females, and low- Bourbou C, Fuller BT, Garvie-Lok SJ, Richards MP. 2011. Reconstructing the diets of Greek Byzantine populations status females show similar dietary trajectories through- (6th–15th centuries AD) using carbon and nitrogen stable iso- out life, it is the low-status males that show different tope ratios. Am J Phys Anthropol 146:569–581. stable isotope values, which suggest a more plant-based Brooks ST, Suchey JM. 1990. Skeletal age determination based diet including millet. The diets of low-status males were on the os pubis: a comparison of the Acsa´di-Nemeske´ri and less predictable and could tend to vary considerably. We Suchey-Brooks methods. Hum Evol 5:227–238.

American Journal of Physical Anthropology DIET AND LIFE HISTORY AT TRINO VERCELLESE 599

Buikstra JE, Ubelaker DH. 1994. Standards for data collection Katzenberg AM. 2000. Stable isotope analysis: a tool for study- from human skeletal remains. Fayetteville, AR: Arkansas Ar- ing past diet, demography, and life history. In: Katzenberg cheological Survey. AM, Saunders SR, editors. Biological anthropology of the Caramiello R, Siniscalco MC, Zeme A, Forlani L, Accorsi CA, human skeleton. New York: Wiley-Liss. Arobba D, Bandini Mazzanti M, Zanini E. 1999. Analisi paleo- Katzenberg AM, Bazaliiskii VI, Goriunova OI, Savel’ev NA, botaniche e sedimentologiche: storia forestale, clima ed agrico- Weber AW. 2010. Diet reconstruction of prehistoric hunter- lutura a Trino dall’eta’ romana al medioevo. Negro Ponzi gatherers in the Lake Baikal region. In: Weber AW, Katzen- Mancini MM, editor San Michele di Trino (VC) Dal villaggio berg AM, Schurr TG, editors. Prehistoric hunter-gatherers of romano al castello medievale, Vol.2. Firenze: All’insegna del the Baikal Region, Siberia. Philadelphia: University of Penn- Giglio. p 577–599. sylvania Museum of Archaeology and Anthropology. Celoria M. 1999. Stress ambientale e condizioni di vita. In: Keenleyside A, Schwarcz H, Panayotova K. 2006. Stable isotopic Mancini MM, editor. San Michele di Trino (VC) Dal villaggio evidence of diet in a Greek colonial population from the Black romano al castello medievale, Vol. 2. Firenze: All’insegna del Sea. J Archaeol Sci 33:1205–1215. Giglio. p 740–749. Kjellstro¨m A, Stora˚ J, Possnert G, Linderholm A. 2009. Dietary Chenery C, Mu¨ldner G, Evans J, Eckardt H, Lewis M. 2010. patterns and social structures in medieval Sigtuna, Sweden, Strontium and stable isotope evidence for diet and mobility in as reflected in stable isotope values in human skeletal Roman Gloucester, UK. J Archaeol Sci 37:150–163. remains. J Archaeol Sci 36:2689–2699. Craig H, Biazzo M, O’Connell TC, Garnsey P, Martinez-Labarga Krueger HW, Sullivan CH. 1984. Models for carbon isotope frac- C, Lelli R, Salvadei L, Tartaglia G, Nava A, Reno L, et al. tionation between diet and bone. In: Turnlund J, Johnson PE, 2009. Stable isotopic evidence for diet at the Imperial Roman editors. Stable isotopes in nutrition. ACS Symposium Series. coastal site of Velia (1st and 2nd centuries AD) in Southern Washington, DC: American Chemical Society. p 205–222. Italy. Am J Phys Anthropol 139:572–583. Le Bras-Goude G, Binder D, Formicola V, Duday H, Couture- Crowe F, Sperduti A, O’Connell TC, Craig OE, Kirsanow K, Ger- Veschambre C, Hublin J-J, Richards MP. 2006. Strate´gies de moni P, Macchiarelli R, Garnsey P, Bondioli L. 2010. Water- subsistance et analyse culturelle de populations ne´olithiques related occupations and diet in two Roman coastal commun- de Ligurie: approche par l’e´tude isotopique (d13Cetd15N) des ities (Italy, first to third century AD): correlation between sta- restes osseux. Bull Me´moires Soc d’Anthropol Paris 18:45–55. ble carbon and nitrogen isotope values and auricular exostosis Linderholm A, Andersson K, Mo¨rth C-M, Grundberg L, Ha˚rding prevalence. Am J Phys Anthropol 142:355–366. B, Lide´n K. 2008a. An early Christian cemetery at Bjo¨rned in Czermak A, Ledderose A, Strott N, Meier T, Grupe G. 2006. northern Sweden. Stable isotope analyses of skeletal material. Social structures and social relations: an archaeological and Fornva¨nnen 103:176–189. anthropological examination of three early medieval separate Linderholm A, Jonson CH, Svensk O, Liden K. 2008b. Diet and burial sites in Bavaria. Anthropol Anz 64:297–310. status in Birka: stable isotopes and grave goods compared. Dembinska M. 1999. Food and drink in medieval Poland: redis- Antiquity 82:446–461. covering a cuisine of the past. Thomas M, translator. Phila- Lovejoy CO, Meindl RS, Pryzbeck TR, Mensforth RP. 1985. delphia: City of Philadelphia Press. Chronological metamorphosis of the auricular surface of the Dyer C. 1994. Everyday life in medieval England. Cambridge: ilium: a new method for determination of age at death. Am J Cambridge University Press. Phys Anthropol 68:15–28. Ervynck A, Van Neer W, Huster-Plogmann H, Schibler J. 2003. McGovern PE, Zhang J, Tang J, Zhang Z, Hall GR, Moreau RA, Beyond affluence: the zooarchaeology of luxury. World Nunez A, Butrym ED, Richards MP, Wang C-S, et al. 2004. Archaeol 34:428–441. Fermented beverages of pre- and proto-historic China. Proc Ferro AM. 1999. La fauna. In: Mancini MM, editor. San Natl Acad Sci Biol 101:17593–17598. Michele di Trino (VC) Dal villaggio romano al castello medie- Minawaga M, Wada E. 1984. Stepwise enrichment of 15N along vale, Vol. 2. Firenze: All’insegna del Giglio. p 631–646. food chains: further evidence and the relation between 15N Fuller BT, Richards MP, Mays S. 2003. Stable carbon and nitro- and animal age. Geochim Cosmochim Acta 48:1135–1140. gen isotope variations in tooth dentine serial sections from Montanari M. 1988. Alimentazione e cultura nel medioevo. Wharram Percy. J Archaeol Sci 30:1673–1684. Rome: Edizioni Laterza. Garvie-Lok SJ. 2001. Loaves and fishes: a stable isotope recon- Moorrees C, Fanning EA, Hunt EJ Jr. 1963. Age variation of struction of diet in medieval Greece [Ph.D. thesis], University formation stages for ten permanent teeth. Int Am Assoc Dent of Calgary, Calgary. Res 42:1490–1502. Girotti M, Garetto TD. 1999. Studio odontologico sulla popola- Mu¨ ldner G, Montgomery J, Cook G, Ellam R, Gledhill A, Lowe zione. In: Mancini MM, editor. San Michele di Trino (VC) C. 2009. Isotopes and individuals: diet and mobility among Dal villaggio romano al castello medievale, Vol. 2. Firenze: the medieval Bishops of Whithorn. Antiquity 83:1119–1133. All’insegna del Giglio. p 732–738. Mu¨ ldner G, Richards MP. 2005. Fast or feast: reconstructing Grupe G, Heinrich D, Peters J. 1999. A brackish water aquatic diet in later medieval England by stable isotope analysis. J foodweb: trophic levels and salinity gradients in the Schlei Archaeol Sci 32:39–48. fjord, Northern Germany, in Viking and medieval times. J Mu¨ ldner G, Richards MP. 2007. Stable isotope evidence for 1500 Archaeol Sci 36:2125–2144. years of human diet at the city of York, UK. Am J Phys Hedges REM, Clement JG, Thomas CDL, O’Connell TC. 2007. Anthropol 133:682–697. Collagen turnover in the adult femoral mid-shaft: modeled Nada Patrone AMN. 1981. Il cibo del ricco ed il cibo del povero: from anthropogenic radiocarbon tracer measurements. Am J contributo alla storia qualitativa dellalimentazione: larea Phys Anthropol 133:808–816. pedemontana negli ultimi secoli del Medio Evo. Turin: Centro Herrscher E. 2003. Diet of a historic population: analysis of iso- studi piemontesi. topic data from the necropolis of Saint-Laurent de Grenoble Negro Ponzi Mancini MM. 1999. San Michele di Trino (VC). Dal (13th to 15th century, France). Bull Memoir Soc d’Anthropol villaggio romano al castello medievale. Firenze: All’insegna Paris 15:149–269. del Giglio. Herrscher E, Bocherens H, Valentin F, Colardelle R. 2001. Com- Nisbet R. 1999. I macroresti vegetali. In: Mancini MM, editor. portements alimentaires au Moyen Aˆ ge a` Grenoble: applica- San Michele di Trino (VC) Dal villaggio romano al castello tion de la bioge´ochimie isotopique a` la ne´cropole Saint-Lau- medievale, Vol. 2. Firenze: All’insegna del Giglio. p 621–630. rent (XIIIe–XVe sie`cles, Ise`re, France). Comptes Rendus Acad Nitsch EK, Hedges REM. 2010. Quantifying changes in fish con- Sci Paris, Sci 324:479–487. sumption in Roman and early medieval Italy using stable iso- Jørkov MLS, Heinemeier J, Lynnerup N. 2009. The petrous tope analysis. Society for Archaeological Sciences 38th Inter- bone: a new sampling site for identifying early dietary pat- national Symposium on Archaeometry, Tampa, FL. terns in stable isotopic studies. Am J Phys Anthropol Nitsch EK, Humphrey LT, Hedges REM. 2011. Using stable iso- 138:199–209. tope analysis to examine the effect of economic change on

American Journal of Physical Anthropology 600 L.J. REITSEMA AND G. VERCELLOTTI

breastfeeding practices in Spitalfields, London, UK. Am J trace element and isotope analyses on early mediaeval skele- Phys Anthropol 146:619–628. tal material. J Archaeol Sci 26:675–685. Phenice TW. 1969. A newly developed visual method of sexing Sealy J, Armstrong R, Schrire C. 1995. Beyond lifetime aver- the os pubis. Am J Phys Anthropol 30:297–302. ages: tracing life histories through isotopic analysis of differ- Polet C, Katzenberg MA. 2003. Reconstruction of the diet in a ent calcified tissues from archaeological human skeletons. An- mediaeval monastic community from the coast of Belgium. J tiquity 69:290–300. Archaeol Sci 30:525–533. Shahar S. 1990. Childhood in the Middle Ages. London: Rout- Porro MA, Boano R, Spani F, Garetto TD. 1999. Studi antropolo- ledge. gici sulla popolazione. In: Mancini MM, editor. San Michele Sˇ laus M. 2008. Osteological and dental markers of health in the di Trino (VC) Dal villaggio romano al castello medievale, Vol. transition from the Late Antique to the Early Medieval period 2. Firenze: All’insegna del Giglio. p 722–732. in Croatia. Am J Phys Anthropol 136:455–469. Privat KL, O’Connell TC, Richards MP. 2002. Stable isotope Spurr MS. 1983. The cultivation of millet in Roman Italy. Pap analysis of human and faunal remains from the Anglo-Saxon Br School Rome 51:1–15. cemetery at Berinsfield, Oxfordshire: dietary and social impli- Steele KW, Daniel RM. 1978. Fractionation of nitrogen isotopes cations. J Archaeol Sci 29:779–790. by animals: a further complication to the use of variations in Prowse T, Schwarcz H, Saunders SR, Macchiarelli R, Bondioli the natural abundance of 15N for tracer studies. J Agric Sci L. 2004. Isotopic paleodiet studies of skeletons from the Impe- 90:7–9. rial Roman-age cemetery of Isola Sacra, Rome, Italy. J Storey R. 1992. Preindustrial urban lifestyle and health. In: Archaeol Sci 31:259–272. Huss-Ashmore R, Schall J, Hediger M, editors. Health and Prowse T, Schwarcz H, Saunders SR, Macchiarelli R, Bondioli lifestyle change. MASCA Research Papers in Science and L. 2005. Isotopic evidence for age-related variation in diet Archaeology. Philadelphia: MASCA, the University Museum from Isola Sacra, Italy. Am J Phys Anthropol 128:2–13. of Archaeology and Anthropology, University of Pennsylvania. Prowse T, Schwarcz H, Saunders SR, Macchiarelli R, Bondioli p 33–42. L. 2007. Isotopic evidence for age-related immigration to Im- Stout SD, Lueck R. 1995. Bone remodeling rates and skeletal perial Rome. Am J Phys Anthropol 132:510–519. maturation in three archaeological populations. Am J Phys Reitsema LJ, Crews DE, Polcyn M. 2010. Preliminary evidence Anthropol 98:161–171. for medieval Polish diet from carbon and nitrogen stable iso- Tafuri MA, Craig OE, Canci A. 2009. Stable isotope evidence for topes. J Archaeol Sci 37:1413–1423. the consumption of millet and other plants in Bronze Age Richards MP, Fuller BT, Molleson TI. 2006. Stable isotope paleo- Italy. Am J Phys Anthropol 139:146–153. dietary study of humans and fauna from the multi-period Tieszen LL, Fagre T. 1993. Effect of diet quality and composi- (Iron Age, Viking, and Late Medieval) site of Newark Bay, tion on the isotopic composition of respiratory CO2, bone col- Orkney. J Archaeol Sci 33:122–131. lagen, bioapatite, and soft tissues. In: Lambert JB, Grupe G, Richards MP, Hedges REM. 1999. Stable isotope evidence for editors. Molecular archaeology of prehistoric human bone. similarities in the types of marine foods used by Late Meso- Berlin: Springer. p 121–155. lithic humans at sites along the Atlantic coast of Europe. J van der Merwe NJ, Vogel JC. 1978. 13C content of human colla- Archaeol Sci 26:717–722. gen as a measure of prehistoric diet in late woodland North Richards MP, Hedges REM, Molleson TI, Vogel JC. 1998. Stable America. Nature 276:815–816. isotope analysis reveals variations in human diet in the Van Neer W, Ervynck A. 2004. Remains of traded fish in Poundbury Camp cemetery site. J Archaeol Sci 25:1247–1252. archaeological sites: indicators of status, or bulk food? In: Richards MP, Mays S, Fuller BT. 2002. Stable carbon and nitro- O’Day SJ, Van Neer W, Ervynck A, editors. Behaviour behind gen isotope values of bone and teeth reflect weaning age at bones: the zooarchaeology of ritual, religion, status and iden- the mediaeval Wharram Percy site, Yorkshire, UK. Am J tity. Oxford: Oxbow Books. p 203–214. Phys Anthropol 119:205–210. Vercellotti G, Stout SD, Boano R, Sciulli P. 2011. Intrapopula- Rutgers LV, van Strydonck M, Bouding M, van der Linde C. tion variation in stature and body proportions: Social status 2009. Stable isotope data from the early Christian catacombs and sex differences in an Italian medieval population (Trino of ancient Rome: new insights into the dietary habits of Vercellese, VC). Am J Phys Anthropol 145:203–214. Rome’s early Christians. J Archaeol Sci 36:1127–1134. Vogel JC, van der Merwe NJ. 1977. Isotopic evidence for early Salamon M, Coppa A, McCormick M, Rubini M, Vargui R, Tur- maize cultivation in New York State. Am Antiq 42:238–242. oss N. 2008. The consilience of historical and isotopic Watt BK, Merrill AL. 1975. Handbook of the nutritional con- approaches in reconstructing the medieval Mediterranean tents of foods. New York: Dover. diet. J Archaeol Sci 35:1667–1672. White L. 1962. Medieval technology and social change. Oxford: Schoeninger MJ. 2011. Diet reconstruction and ecology using Oxford University Press. stable isotope ratios. In: Larsen CS, editor. A companion to Woolgar C. 2000. ‘‘Take this penance now, and afterwards the biological anthropology. Chichester: Wiley-Blackwell. fare will improve’’: seafood and late medieval diet. In: Star- Schoeninger MJ, DeNiro MJ. 1984. Nitrogen and carbon isotopic key DJ, Reid C, Ashcroft N, editors. England’s sea fisheries: composition of bone collagen from marine and terrestrial ani- the commercial sea fisheries of England and Wales since mals. Geochim Cosmochim Acta 48:625–639. 1300. London: Chatham. p 36–44. Schutkowski H, Herrmann B, Wiedemann F, Bocherens H, Zar JH. 1999. Biostatistical analysis. Upper Saddle River, NJ: Grupe G. 1999. Diet, status and decomposition at Weingarten: Prentice Hall.

American Journal of Physical Anthropology Table 2. Human stable isotope and collagen quality results. Dentine Bone Sample ID Sex Status %N %C C:N δ15N (‰) δ13C (‰) %N %C C:N δ15N (‰) δ13C (‰) S-46 Female High 15.2 41.7 3.2 9.0 -19.8 14.2 41.6 3.4 9.1 -19.8 S-48A Female High 14.4 40.0 3.2 9.2 -18.5 13.8 40.2 3.4 10.5 -19.4 S-73 Female High 14.5 40.4 3.2 9.2 -19.9 14.4 41.8 3.4 9.4 -19.9 S-80 Female High 15.0 41.4 3.2 9.2 -18.7 14.2 40.9 3.4 9.2 -19.7 S-207 Female Low 15.3 42.2 3.2 9.6 -20.1 14.2 40.0 3.3 9.4 -19.8 S-297 Female Low 15.3 42.0 3.2 8.4 -18.6 14.8 41.3 3.3 9.0 -19.3 S-348 Female Low 15.1 41.4 3.2 9.4 -19.4 14.6 40.6 3.2 9.1 -19.3 S-528 Female Low 15.5 42.9 3.2 6.7 -20.0 13.9 39.6 3.3 9.1 -19.1 S-535 Female Low 15.1 41.4 3.2 9.7 -19.4 10.2 32.4 3.7 8.3 -19.1 S-542 Female Low 15.7 43.0 3.2 10.3 -18.7 14.8 41.2 3.3 9.3 -19.5 S-41 Male High 15.2 42.5 3.3 10.4 -18.6 12.2 35.8 3.4 10.3 -19.9 S-44 Male High 14.5 40.5 3.3 9.9 -19.0 12.3 35.5 3.4 9.7 -19.6 S-56 Male High 15.6 43.3 3.3 9.5 -19.6 5.0 15.3 3.5 9.0 -19.2 S-57 Male High 15.0 41.7 3.2 9.5 -19.5 3.7 11.1 3.5 8.9 -19.4 S-68 Male High 14.9 41.2 3.2 8.9 -19.4 11.8 34.3 3.4 8.7 -19.7 S-133 Male High 14.9 41.3 3.2 9.1 -19.6 12.5 35.5 3.3 10.0 -19.0 S-134 Male High 15.2 42.2 3.2 9.1 -19.6 2.8 8.4 3.5 8.6 -19.3 S-143 Male High 15.6 43.0 3.2 9.1 -18.2 7.3 20.7 3.3 8.8 -18.5 S-166-1 Male High 14.6 40.8 3.3 9.3 -19.5 12.9 36.8 3.3 9.8 -18.9 S-166-2B Male High 13.1 39.1 3.5 9.5 -19.5 7.8 22.2 3.3 9.4 -18.9 S-328 Male Low 15.9 43.7 3.2 10.8 -19.8 6.8 19.0 3.3 8.1 -17.8 S-346 Male Low 14.7 41.0 3.2 9.6 -18.5 5.6 16.1 3.3 8.4 -18.5 S-367 Male Low 15.1 41.8 3.2 9.8 -18.3 7.7 21.3 3.3 8.8 -17.9 S-398 Male Low 15.1 41.8 3.2 8.4 -19.9 12.8 35.4 3.2 8.6 -17.4 S-408 Male Low 15.3 42.5 3.2 12.0 -20.0 11.9 33.4 3.3 11.8 -19.3 S-424 Male Low 14.9 41.6 3.3 9.0 -19.4 0.7 3.7 6.1 2.00 -20.9 S-456 Male Low 15.3 42.4 3.2 8.3 -17.8 9.7 26.9 3.2 8.3 -18.9 S-474 Male Low 14.7 41.4 3.3 9.8 -19.1 12.7 34.5 3.2 9.7 -19.1 S-545 Male Low 15.5 42.7 3.2 9.5 -17.6 8.4 23.6 3.3 8.1 -17.9 S-555 Male Low 15.2 42.2 3.2 8.0 -19.1 12.5 34.0 3.2 8.1 -18.6